arxiv:0910.4262v1 [astro-ph.sr] 22 oct 2009
TRANSCRIPT
arX
iv:0
910.
4262
v1 [
astr
o-ph
.SR
] 2
2 O
ct 2
009
The Clusters AgeS Experiment (CASE). IV.
Analysis of the Eclipsing Binary V69 in the Globular Cluster
47 Tuc1
I. B. Thompson2, J. Kaluzny3, S. M. Rucinski4, W. Krzeminski5, W. Pych3, A. Dotter6,
G. S. Burley2
ABSTRACT
We use photometric and spectroscopic observations of the eclipsing binary
V69-47 Tuc to derive the masses, radii, and luminosities of the component stars.
Based on measured systemic velocity, distance, and proper motion, the system
is a member of the globular cluster 47 Tuc. The system has an orbital pe-
riod of 29.5 d and the orbit is slightly eccentric with e = 0.056. We obtain
Mp = 0.8762 ± 0.0048M⊙, Rp = 1.3148 ± 0.0051R⊙, Lp = 1.94 ± 0.21L⊙
for the primary and Ms = 0.8588 ± 0.0060M⊙, Rs = 1.1616 ± 0.0062R⊙,
Ls = 1.53 ± 0.17L⊙ for the secondary. These components of V69 are the
first Population II stars with masses and radii derived directly and with an
accuracy of better than 1%. We measure an apparent distance modulus of
(m−M)V = 13.35 ± 0.08 to V69. We compare the absolute parameters of V69
with five sets of stellar evolution models and estimate the age of V69 using mass-
luminosity-age, mass-radius-age, and turnoff mass - age relations. The masses,
radii, and luminosities of the component stars are determined well enough that
the measurement of ages is dominated by systematic differences between the
evolutionary models, in particular, the adopted helium abundance. By com-
paring the observations to Dartmouth model isochrones we estimate the age of
V69 to be 11.25±0.21(random)±0.85(systematic) Gyr assuming [Fe/H] = -0.70,
2Carnegie Observatories, 813 Santa Barbara St., Pasadena, CA 91101-1292;
(ian,burley)@obs.carnegiescience.edu
3Copernicus Astronomical Center, Bartycka 18, 00-716 Warsaw, Poland; (jka,pych)@camk.edu.pl
4David Dunlap Observatory, Department of Astronomy and Astrophysics, University of Toronto, P.O.
Box 360, Richmond Hill, ON L4C 4Y6, Canada; [email protected]
5Las Campanas Observatory, Casilla 601, La Serena, Chile; [email protected]
6Department of Physics and Astronomy, University of Victoria, P.O. Box 3055, Victoria, B.C., V8W 3P6,
Canada; [email protected]
– 2 –
[α/Fe] = 0.4, and Y = 0.255. The determination of the distance to V69, and
hence to 47 Tuc, can be further improved when infrared eclipse photometry is
obtained for the variable.
Subject headings: binaries: close – binaries: spectroscopic – stars: individual
V69 47 Tuc – globular clusters: individual (47 Tuc)
1. INTRODUCTION
Detached eclipsing double-line binary stars are the fundamental astrophysical laboratory
for the determination of stellar parameters of mass and radius. Luminosities can be derived
using measured parallaxes or from empirical color – effective temperature relations. These
data are the fundamental tests of stellar evolution models. Many field Population I systems
are known at solar mass and larger (Andersen 1991) and modern high accuracy measurements
of masses, luminosities, and radii of the component stars are in general agreement with
evolution models (see, for example, Lacy et al. 2005, 2008; Clausen et al. 2008). Similar
results are obtained for studies of individual binaries in the old open clusters NGC 188
(Meibom et al. 2009), NGC 2243 (Kaluzny et al. 2006), and NGC 6791 (Grundahl et al.
2008). Recent wide field photometric surveys have identified numerous low mass systems,
and for these K and M stars the common theme of a comparison of component properties with
evolution models is that the models systematically underestimate the radii of the components
in these binaries. Summaries of recent measurements can be found in Lopez-Morales et al.
(2006), Lopez-Morales (2007), and Blake et al. (2008).
The situation is even less clear for Population II stars. With the exception of CM Dra,
for which the component masses are ∼0.2M⊙ (Lacy 1977), and the ω Cen binary OGLEGC-
V17, for which the analysis is compromised by an uncertain determination of the metallicity
(Thompson et al. 2001; Kaluzny et al. 2002), there are no known Population II detached
double-line eclipsing binaries with main-sequence components. Torres et al. (2002) used
interferometric observations of HD 195987 ([Fe/H] ∼ -0.6) to derive an orbit and measure the
masses of the components. The radiative properties of the two components agree with a suite
of models given some slight modifications of input parameters. While direct measurements
of the radii of the components are not possible because HD 195987 is not an eclipsing binary,
1This paper includes data gathered with the 6.5-meter Magellan Baade and Clay Telescopes and the 2.5-
meter du Pont Telescope located at Las Campanas Observatory, Chile. It is based in part on data obtained
at the South African Astronomical Observatory.
– 3 –
estimates of the radii can be derived from the orbital parallax, the bolometric flux, and the
estimated effective temperatures. Here again the measured radii are larger than the models
by some ten per cent. Boyajian et al (2008) have measured the angular diameter of the G
subdwarf µ Cas ([Fe/H ∼ -0.8). This measurement provides a radius when combined with
the Hipparchos parallax for µ Cas. For this star the models underpredict this measured
radius by about five per cent. The masses of the components of µ Cas are only known to
about ten per cent, so the model comparisons here are not well constrained.
There is a clear need to locate and study Population II detached eclipsing binary stars
to obtain accurate masses and radii of their component stars. Stellar evolution models
are becoming increasingly sophisticated and are used to fit observed cluster color-magnitude
diagrams (CMD). The cluster CMD’s are themselves improving in quality, with homogeneous
surveys being conducted with the Hubble Space Telescope (see, for example, Sarajdeni et al.
2007). Careful empirical tests of these models are an essential next step.
This is the first paper in a series devoted to the study of detached eclipsing double-line
binaries (DEB) in Galactic globular clusters with components on the cluster main sequence
or subgiant branch. The Cluster AgeS Experiment (CASE) has the goal of determining
the basic stellar parameters (masses, luminosities, and radii) of the components of cluster
binaries to a precision of better than 1% in order to measure cluster ages and distances,
and to test stellar evolution models. The methods and assumptions utilize basic and simple
approaches offered by the field of eclipsing double-line spectroscopic binaries as described in
Paczynski (1997) and Thompson et al. (2001). Previous CASE papers have discussed blue
straggler systems in ω Cen (Kaluzny et al. 2007a) and 47 Tuc (Kaluzny et al. 2007b), and
an SB1 binary in NGC 6397 Kaluzny et al. (2008).
The eclipsing binary V69-47 Tuc (hereinafter V69) was discovered by Weldrake et al.
(2004) during a survey for variable stars in the field of the globular cluster 47 Tuc. They
presented an I-band light curve for the variable and proposed an orbital period of P =
5.229 d. The light curve phased with this period shows only one eclipse. The variable is
located at the top of main sequence in the cluster color-magnitude diagram, and thus is of
potential great interest for measurements of the cluster age and distance.
In this paper we report the results of photometric and spectroscopic observations aimed
at a determination of the absolute parameters of the components of V69. Section 2 describes
the photometry of the variable and the determination of an orbital ephemeris. Section 3
presents the radial velocity observations. The combined photometric and spectroscopic ele-
ment solutions are given in Section 4 while the membership in 47 Tuc is discussed in Section 5.
In Section 6 we compare the properties of the components of V69 to a selection of stellar
evolution models with an emphasis on estimating the age of the system. Finally in Section 7
– 4 –
we summarize our findings.
2. PHOTOMETRIC OBSERVATIONS
The bulk of the photometric data were obtained with the 1.0-m Swope telescope at the
Las Campanas Observatory using the 2048 × 3150 pixel SITE3 CCD camera with a scale
of 0.435 arcsec/pixel. These observations were collected during the 2004 – 2007 observing
seasons. The same set of BV filters was used for all observations. Exposure times ranged
from 100 s to 360 s for the V filter (average exposure was 120 s) and from 160 s to 360 s
for the B filter (average exposure was 170 s). An eclipse event was detected on the first
of a total of 8 nights of observations during the 2004 season. When the 2004 season data
were combined with the observations of Weldrake et al. (2004) we were able to eliminate
the proposed 5.25-d period, but the combined data set was insufficient to establish a unique
ephemeris. We then examined the 286 V band images collected by the OGLE team on 44
nights during the 1993 observing season for the cluster field 104-A (Kaluzny et al. 1998). A
single eclipse event was detected on the night of 1993 July 22 UT and this permitted the
identification of an approximate orbital period of P ≈ 29.540 d. Subsequent observations
collected in the 2005, 2006, and 2007 seasons concentrated on nights with predicted eclipse
events. During the 2007 season we observed the variable with the 2.5-m du Pont telescope
using the 2048 × 2048 pixel TEK5 CCD camera at a scale of 0.259′′/pixel. These data
included one well covered primary eclipse observed mostly in the V -band on 2007 August 18
UT.
In addition to the Las Campanas photometry, we obtained V -band observations of an
eclipse on 2005 October 22 UT using the 1.0-m telescope at the South African Astronomical
Observatory. Observations were obtained with the STE4 1024 × 1024 pixel CCD camera at
a scale of 0.31′′/pixel.
Profile photometry was extracted for all of the observations using the DAOPHOT/ALLSTAR
package (Stetson 1987). We calibrated the instrumental photometry using observations col-
lected with the du Pont telescope on the night of 2007 August 19 UT. Observations were
made of V69 together with 4 Landolt standard fields (Landolt 1992). The standards were
observed with a range of air-mass of 1.19 < X < 1.94. The conditions were photometric and
the seeing ranged from 1.05 to 1.40 arcsec with a median value of 1.22 arcsec. The images of
the cluster itself were obtained at an air-mass of 1.38 with seeing of 1.0 arcsec. Profile pho-
tometry for the field of the variable was extracted from subframes covering 220× 180 arcsec
(the full field of TEK5 camera is 8.65 × 8.65 arcmin). This helped to minimize the effects
of a variable point-spread function and as a result to obtain reliable aperture corrections.
– 5 –
Aperture corrections for the V69 observations and the standard field observations were de-
rived using the program DAOGROW (Stetson 1990). Magnitudes of 28 standard stars in the
Landolt fields were taken from the Stetson catalog (Stetson 2000)2. The following relations
between instrumental (lower case letters) and standard magnitudes were obtained:
v = V − 0.026(2) × (B − V ) + 0.118(3) ×X + const, (1)
b = B − 0.069(2) × (B − V ) + 0.217(3) ×X + const, (2)
where X is the air-mass. In Fig. 1 we show the residuals between the standard and recovered
magnitudes for the Landolt primary standards. The analysed V69 field includes 20 additional
secondary standard stars from the Stetson catalog (Stetson 2000). The average residuals for
these stars are ∆V = +0.004± 0.019 and ∆(B − V ) = −0.001 ± 0.015 with our magnitudes
fainter and our colors bluer on average.
A total of 8 primary and 6 secondary eclipses were observed in the combined data sets.
In Fig. 2 we show the BV light curves of V69 phased with the ephemeris :
MinI = HJD 2453237.8421(2) + 29.53975(1) (3)
The period was derived using the Lafler-Kinman algorithm, and the error in the period was
estimated by visual inspection of the phased light curve as the period was varied away from
the best fit value. These light curves contain a total of 1216 and 310 data points for V
and B, respectively. The plots include photometry from all four data sets mentioned above.
For the OGLE data we have plotted only points inside the primary eclipse. The colors and
magnitudes of V69 at minima and at quadrature are listed in Table 1. The quoted errors do
not include possible systematic errors of the zero points of the photometric solution which
we estimate to be 0.010 mag.
3. SPECTROSCOPIC OBSERVATIONS
We obtained spectroscopic observations of V69 with the MIKE echelle spectrograph
(Bernstein et al. 2003) on the Magellan Clay 6.5-m telescope. All observations were taken
with a 0.7 arcsec slit at a resolution of R ≃ 40,000. The observations generally consisted of
two exposures flanking an exposure of a thorium-argon hollow-cathode lamp. Total exposure
times per spectrum ranged from 1300 to 3665 seconds depending on observing conditions.
2We have used the electronic version of the catalog as of 2007 Novem-
ber 7. The catalog is maintained by Canadian Astronomy Data Centre at
http://www3.cadc-ccda.hia-iha.nrc-cnrc.gc.ca/community/STETSON/standards/
– 6 –
The observations were reduced with pipeline software written by Dan Kelson following the
approach of Kelson (2003). Post-extraction processing of the spectra was done within the
IRAF ECHELLE package.3
Velocities were measured with the TODCOR algorithm (Zucker & Mazeh 1994) using
an implementation written by G. Torres. For velocity templates we used synthetic echelle
resolution spectra from the grid of Coelho et al. (2006). We adopted the values of log g and
Teff derived from the photometric solution (see next section) and assumed a metallicity of
[Fe/H] = −0.71 with an α-element enhancement of 0.4 (see the discussion of the metallicity
of 47 Tuc in the next section). The templates were adjusted once during the iterations
for the Wilson-Devinney solutions, and final linear interpolations on the Coelho et al. grid
were made at (log g, Teff and [Fe/H]) = (4.14, 5945 K, −0.71) for the primary and (4.24,
5955 K, −0.71) for the secondary. The measured velocities are insensitive to minor changes
in these parameters. The interpolated synthetic spectra were smoothed with a gaussian
to match the resolution of the observations. No rotational broadening was applied to the
templates. The cross-correlations covered the wavelength intervals 4125 A < λ < 4320 A,
4350 A < λ < 4600 A, and 4600 A < λ < 4850 A. The final adopted velocities are the
averages of these three measurements for each of the observations.
The measured radial velocities were fit with a non-linear least squares solution using
code written by G. Torres adopting the ephemeris given in equation (3). The observations
are presented in Table 2 which lists the heliocentric Julian Date (HJD) at mid-exposure,
the velocities of the primary and secondary components, and the orbital phases of the ob-
servations. The adopted orbital elements are listed in Table 3 and the orbit is plotted in
Fig. 3.
4. LIGHT CURVE ANALYSIS AND SYSTEM PARAMETERS
We have used two different models for the analysis of the light curves. The first
is the Wilson-Devinney model (Wilson & Devinney 1971; Wilson 1979) as implemented
in the PHOEBE package (Prsa & Zwitter 2005). The second is the JKTEBOB program
(Southworth et al. 2004a,b), which is based on the EBOP code (Popper 1980; Popper & Etzel
1981; Etzel 1981). The most recent public version of JKTEBOB is described in detail in
Southworth et al. (2007). In particular, it incorporates an option to adopt a non-linear limb
darkening law.
3IRAF is distributed by the National Optical Astronomy Observatories, which are operated by the Asso-
ciation of Universities for Research in Astronomy, Inc., under cooperative agreement with the NSF.
– 7 –
The two light curve fitting programs utilize different methodologies and approaches.
The PHOEBE/WD program is more sophisticated in its treatement of the geometry and
the stellar atmospheres. It utilizes Roche geometry to approximate the shapes of the stars,
uses Kurucz model atmospheres, and treats the reflection effect in detail. It can be used
for any component separation including contact systems. JKTEBOP/EBOP approximates
stars by bi-axial ellipsoids. It is faster than the WD code but is appropriate only for binary
components that are spherical or only slightly distorted. The real advantage of JKTEBOP is
that the errors for the fit parameters are reliably determined. The PHOEBE code provides
only formal errors of the fits and these are known to be underestimated (A. Prsa, personal
communication). However an advantage of the PHOEBE code is that it can be used for the
analysis of multi-color light curves. For V69 the solution for the B-band is much poorer than
for the V -band because of the lower quality of the B-band data. PHOEBE/WD was used
to simultaneously fit the V and B curves. The resulting geometrical parameters are based
mostly on the higher quality V curve while the simultaneous fit permits a good estimate of
the luminosity ratio.
In the analysis we assumed that the light curve of V69 is free from any “third light”.
This is supported by the depths of the eclipses, with ∆V > 0.63 mag for both the primary
and secondary eclipse. There is also no evidence for any third component in the spectra of
the binary. Finally, we note that V69 can be located on numerous HST/ACS images (for
example, PEP ID 9018, P.I. G. De Marchi). Examination of these images indicates that
our ground-based photometry does not suffer from any blending problems from unresolved
visual companions to the variable.
Four recent high resolution studies of the metallicity of 47 Tuc suggest a value of
[Fe/H] ∼ −0.7. As part of a study to establish a globular cluster metallicity scale based
on measurements of Fe II lines, Kraft & Ivans (2003) reanalysed in a uniform way equivalent
widths in a total of 8 giants measured by Brown & Wallerstein (1992), Norris & Da Costa
(1995), and Carretta & Gratton (1997). They obtained a range of -0.61 to -0.69 for [Fe/H]Iand a range of -0.56 to -0.70 for [Fe/H]II. Alves-Brito et al. (2005) used observations of
5 giants to obtain [Fe/H]I = -0.66 and [Fe/H]II = -0.69 with an α-element enhancement
of about 0.3 dex. Koch & McWilliam (2008) found [Fe/H]I = -0.76 and [Fe/H]II = -0.82
with an α-element enhancement of 0.4 dex based on a study of eight giants and one dwarf
star. Finally, Carretta et al. (2009) measured [Fe/H]I = -0.77 and [Fe/H]II = -0.80 with
an α-element enhancement of +0.39 from observations of 11 giants. For this analysis, we
have adopted a value of [Fe/H] = -0.71 together with an α-element enhancement of +0.4.
Sensitivities of age estimates to these assumptions are explored in Sections 6.1 and 6.2.
The effective temperature of the primary is needed for the PHOEBE light curve solution
– 8 –
and for both components to estimate the luminosities of the stars from the derived radii. In
the absence of detailed infrared eclipse photometry we estimated the effective temperature
of the primary, Tp, from the dereddened (B − V ) and (V − I) color indices. We used the
observed color of the binary, (B − V ) = 0.548 (see Table 1) together with a reddening of
E(B − V ) = 0.04 (Harris 1996; Gratton et al. 2003; Percival et al. 2002) to derive (B − V )0= 0.508. For the (V − I) color we note that V69 is star number 14065 in the catalog of
Kaluzny et al. (1998) where (V − I) = 0.731±0.025. We shift this observed color by -0.026
to put the observation on the system of Stetson (2000) (Percival et al. 2002) and adopt the
reddening law of Schlegel et al. (1998) to obtain E(V −I) = 0.06 and a final value of (V −I)0= 0.671. The similarity of the (B − V ) colors in and out of eclipse (see Table 1) suggests
that the effective temperature of the primary derived from these colors is close to the true
value.
Visual inspection of the position of the binary in a color-magnitude diagram suggests
that the components are beginning to evolve away from the main sequence (see Fig. 4), and
so we adopt an initial gravity for the primary of log g = 4.1.
These adopted colors and parameters lead to estimates for the effective temperature of
the primary of 6070 K and 5900 K, 6050 K and 5905 K, and 5960 K and 5780 K for the (B−V )
and (V − I) colors and the calibrations of Worthey & Lee (2006), Ramirez & Melendez
(2005), and VandenBerg & Clem (2003), respectively. We adopt a linear average of these
estimates, resulting in an effective temperature for the primary of 5945 ± 105 K, where the
quoted error is the standard deviation of the six temperature estimates.
The three color-temperature relations considered here do not explicitly account for α-
element enhancement. Based on the synthetic color-temperature employed by Dotter et al.
(2007), we estimate that the difference in Teff between [α/Fe]=0 and +0.4 is ∼10 K for
the [Fe/H], Teff , and log g of the primary. We expect this is an upper limit because the
color-temperature relations should include some implicit dependence on [α/Fe]. In any case,
the [α/Fe]-color uncertainty is small compared to that which arises between the different
color-temperature relations or between B − V and V − I.
We note that Koch & McWilliam (2008) adopted an effective temperature of 5750 K
for the turnoff star that they studied based on excitation equilibrium for Fe I lines, and it is
of interest to ask how accurate our photometric estimate of the effective temperature of the
primary of V69 actually is. We estimate the uncertainty in the reddening to be 0.010 mag,
and this value coupled with our estimated uncertainty in the photometric calibration sug-
gests that the uncertainty in the derived unreddened (B − V ) colors of the components
of V69 to be ∼0.014 mag. The (V − I) color is somewhat less accurate. For a range of
metallicity of [Fe/H] from −0.61 to −0.77 (see above), an estimated range in the gravity of
– 9 –
0.2 dex and a uncertainty in (B − V )0 of 0.014, the Worthey-Lee calibration suggests an
uncertainty in the derived effective temperature of approximately 80 K. The calibrations of
Ramirez & Melendez (2005), and VandenBerg & Clem (2003) indicate similar uncertainties.
Considering the scatter about Teff in the various calibrations we conservatively conclude
that we can estimate the effective temperature of the components of V69 to an accuracy
of about 150 K. These three calibrations predict bolometric corrections spanning the range
−0.11 to −0.09 and we adopt BCp = BCs = −0.10 ± 0.01.
4.1. Models Using the PHOEBE Code
The PHOEBE implementation of the Wilson-Devinney model fits the orbital inclination
i, the gravitational potentials Ωp and Ωs, the effective temperature of the secondary Ts, the
eccentricity e, the longitude of the periastron ω and the relative luminosities Ls/Lp in B
and V simultaneously. We iterated on the solution, calculating subsequent values for the
(B − V )0 colors for the primary and secondary from the observed (B − V )0 for the system
and the modeled values for (Ls/Lp)B and (Ls/Lp)V at each iteration. The mass ratio was set
to the spectroscopic value of q = masss/massp = 0.984. The iterations converged in two
cycles. The final values of the fit are listed in Table 4. The values for the relative radii are
derived directly from the non-dimensional potentials Ω1 and Ω2, and the spectroscopic mass
ratio q. The luminosity ratios and the relative radii are insensitive to changes in effective
temperature: (Ls/Lp)B, (Ls/Lp)V , rs, and rp all change by less than 0.3 per cent for a
±150 K change in effective temperature.
4.2. Models Using the JKTEBOB Code
The JKTEBOB code optimizes the sum of the relative radii rp + rs, the ratio k = rs/rp,
the ratio of the surface brightness of the two stars J , the orbital eccentricity e, and the
longitude of the periastron ω. These two last parameters were included in the analysis
by fitting e cos ω and e sin ω. The mass ratio was fixed at the spectroscopic value of
q = 0.984, and the gravity brightening exponent was set to 0.32. Note that the values for
these two parameters have a negligible effect on the model light curve as the stars are very
well separated and both are practically spherical. We adopted a square-root law for the limb
darkening and adopted theoretical limb darkening coefficients from van Hamme (1993) for
Teff = 5959 K (adopted from the PHOEBE code solution) and [Fe/H] = −0.71.
Table 5 contains the values of the fit parameters along with other relevant parameters
– 10 –
derived from the solution. The last two rows list the reduced chi-squared of the fit and the
rms of the residuals. Separate solutions were derived for light curves in the V and B bands.
The uncertainties of individual parameters were estimated using Monte Carlo simulations
as described in Southworth et al. (2004b). Ten thousand simulations were run for each light
curve. The solution based on the V light curve is much better constrained than the solution
for the B band. The residuals of the fit obtained with the JKTEBOB are shown in Fig. 5.
4.3. Adopted Stellar Parameters
A comparison of Tables 4 and 5 shows that the two light curve synthesis codes give very
similar results. There is also good agreement between the fitted values of e and ω and those
derived from the spectroscopic observations. The remarkably small errors for the fit radii
and relative luminosities might seem suspicious at first given that the light curves suggest
that the eclipses of V69 are partial. However, the secondary eclipse is very close to being
total. Examination of the synthetic light curves shows that only 0.4% of the surface of the
secondary component remains visible at the center of the secondary eclipse.
The magnitudes for the individual components of the binary are derived from the lumi-
nosity ratios from the light curve solution and the observed out-of-eclipse photometry. We
obtain Vp = 17.468±0.010, Bp = 18.019±0.010, Vs = 17.724±0.010 and Bs = 18.268±0.010
where the errors are dominated by the contribution from zero point uncertainties of our pho-
tometry, estimated at 0.010 mag. The colors of both components are practically identical
with (B− V )p = 0.551± 0.014 and (B− V )s = 0.544± 0.014. The position of the binary on
the 47 Tuc color – magnitude diagram is shown in Fig. 4. The primary component has left
the main-sequence of the cluster and is starting to ascend onto the subgiant branch. The
secondary also has begun to evolve and is among the bluest stars at the cluster turnoff.
The absolute parameters of V69 obtained from our spectroscopic and photometric anal-
ysis are given in Table 6. Because the JKTEBOP code provides better error estimates than
PHOEBE we have adopted the V -band relative radii from Table 5. Note, however, that
the relative radii derived with the PHOEBE models agree at a level of 0.2% with those ob-
tained with the JKTEBOP models. The luminosities of the components follow directly from
log(L/L⊙) = 2 log(R/R⊙) + 4 log(Teff/T⊙) where we have adopted R⊙ = 6.9598 × 105 km
(Bahcall et al. 2005), and T⊙ = 5777K (Neckel 1986).
– 11 –
5. MEMBERSHIP OF V69 AND THE DISTANCE OF 47 TUC
Before using the derived stellar parameters of the components of V69 to estimate the
age of 47 Tuc, it is appropriate to consider the cluster membership of the star. There
are four lines of evidence available. First, although the velocity of 47 Tuc is low at vrad =
−18.7±0.5 km sec−1 (Gebhardt et al. 1995), and thus does not provide a strong discriminant
against projected field stars, the center-of-mass velocity of V69 (γ = −16.36 km sec−1) is
consistent with this velocity. At the location of the variable – about 6 arcmin from the
cluster center – the velocity dispersion of cluster stars is about 8 km sec−1. Second, Fig. 4
shows that the individual components of V69 lie on the cluster main sequence. Third,
the binary is a proper motion member of 47 Tuc. The absolute proper motion of 47 Tuc
has been measured to be µα = 5.64± 0.20 mas/yr and µδ = − 2.05± 0.20 mas/yr
(Anderson & King 2003). V69 is located in Field F of that study. This field has subsequently
become a standard calibration field for HST, and V69 is present in numerous ACS/WFC
frames taken between 2002 and 2006. A quick study of the ∼30s F606W exposures shows that
the proper motion of V69 relative to the bulk of the cluster stars is 0.2 mas/yr in RA and -0.1
mas/yr in Dec, a typical internal motion for a cluster member at this radius (J. Anderson,
personal communication). Finally, based on the absolute magnitudes listed in Table 6 and
the apparent magnitudes listed in Sec. 4.3 we derive distance moduli for the components of
V69 of (m−M)V = 13.35±0.11 and (m−M)V = 13.34±0.12 for the primary and secondary,
respectively. The average of these values is (m−M)V = 13.35±0.08. We compare this value
to some recent distance determinations for 47 Tuc in Table 7. While Bono et al. (2008) have
shown that there appears to be some systematic errors in the measurement of the distance
to 47 Tuc by different techniques, V69 clearly lies at the distance of 47 Tuc. We note in
particular that Kaluzny et al. (2007b) find a distance modulus of (m−M)V = 13.40 ± 0.08
for OGLE-228, another eclipsing binary in 47 Tuc. We conclude that V69 is a member of
the globular cluster 47 Tuc.
6. THE AGE OF 47 TUC
The potential of using observations of eclipsing binaries for a robust determination of
globular cluster ages was advocated by Paczynski (1997). In particular, he argued that
the mass-luminosity-age relation be used rather than the mass-radius-age relation since the
latter can be affected by inaccuracies in the stellar models related to the treatment of sub-
photospheric convection. In the following we compare the properties of the components of
V69 with five different sets of stellar evolution models. We have attempted to minimize the
effects of metallicity and α-element enhancement by comparing models with uniform values
– 12 –
of [Fe/H] = -0.71 and, where possible, [α/Fe] = +0.4.
In our analysis we use the Dartmouth Stellar Evolution Database (Dotter et al. 2007),
the Padova models (Salasnich et al. 2000), the Teramo models (Pietrinferni et al. 2006), the
Victoria-Regina models (VandenBerg et al. 2006), and the Yonsei-Yale models (Kim et al.
2002).4 Table 8 compares the parameters of each model set used in the derivation of the ages
of V69. Entries for the solar abundance scale refer to Grevesse & Noels (1993) (Z/X ∼ 0.024))
and Grevesse & Sauval (1998) (Z/X ∼ 0.023). We note that solar models based on recent
measurements of the solar abundance scale remain in conflict with helioseismology data (see
Serenelli et al. (2009) for a discussion). We compare the derived values of mass, luminosity,
and radius for the components of V69 with the isochrones. For the Dartmouth, Victoria-
Regina, and Yonsei-Yale models we used interpolation software provided by the different
groups to generate the plotted isochrones. For the Padova models we linearly interpolated
on log (age) to generate isochrones spaced by 0.5 Gyr for ease of comparison with the other
models.
We plot the components of V69 on the mass-luminosity and mass-radius planes for the
Dartmouth, Padova, Teramo, Victoria-Regina, and Yonsei-Yale models in Fig. 6, Fig. 7,
Fig. 8, Fig. 9, and Fig. 10, respectively. In each Figure isochrones are shown with a step
of 0.5 Gyr. The measured ages for the components of V69 are summarized in Table 9. To
calculate the random errors in the age estimates we assumed that the measurement errors
in mass, luminosity and radius are uncorrelated. This is a reasonable assumption for the
errors in luminosity are dominated by the adopted error in Teff and the errors in radius
are dominated by the photometric solution rather than the error in A sin i as measured by
the orbital solution. The errors in mass, luminosity, and radius lead to uncertainties in age
through comparison with the sets of isochrones, and the final random errors listed in Table 9
are the quadrature sums of the mass-radius and mass-luminosity age uncertainties. The
uncertainties in the masses (about 0.6%) and in the luminosities (about 10%) give similar
contributions to the errors of the derived ages, while the random errors in the age estimated
from mass-radius-age relations are completely dominated by the measurement errors in the
masses of the components. We emphasize that the plotted uncertainty in luminosity arises
directly from the ±150 K uncertainty in the effective temperatures of the components of V69.
The radii are measured with a better relative accuracy than the luminosities and therefore
4The models and supporting software are available at the following websites: Dart-
mouth (http://stellar.dartmouth.edu/∼models/), Padova (http://pleiadi.oapd.inaf.it/),
Teramo (http://193.204.1.62/index.html), Victoria-Regina (http:www.cadc-
ccda.hia-iha.nrccnrc.gc.ca/cvo/community/VictoriaReginaModels/), and Yonsei-Yale
(http://www.astro.yale.edu/demarque/yyiso.html).
– 13 –
the ages can be derived with a random error a factor of two smaller than the ages derived
from the mass-luminosity relations. Table 9 also presents the weighted average of the ages
measured for the primary and secondary components of the binary.
Finally, since the components are very close to the cluster turnoff (see Fig. 4), the
age of the system can be estimated from turnoff mass-age relations. For each isochrone we
estimated the turnoff mass as the mass at maximum log Teff in that isochrone. Based on
the position of the components in the CMD we conclude that the secondary is at the blue
extreme of the CMD, and that the primary has begun to evolve onto the subgiant branch
at a slightly cooler effective temperature. Our mass measurements are accurate enough to
resolve the progression of mass along an isochrone in this part of the CMD. Fig. 11 shows the
resulting relations, and these derived ages are also summarized in Table 9. For completeness,
ages are also given for the primary assuming it is at the cluster turnoff. The quoted errors
in these ages are derived assuming only an error in the measurement of the masses, and are
measured from the values of the ages at the ±1σ extrema of the measured masses. This age
measurement is obviously fairly crude, but serves as a consistency check on the other ages
estimates.
The ages of V69 determined from these relations show some considerable spread from
one model set to the next, ranging from 11.1 to 13.7 Gyr. While we only have age estimates
from two stars, we can draw some general conclusions. The first is that the close agreement
in the age estimates for the primary and secondary for all models suggests that the errors in
luminosity are overestimated. Second, the ages measured by the three methods also agree to
within the random errors within each model. This suggests that concerns about the model
radii arising from uncertainties in the treatment of convection are overstated, and that the
models predict internally consistent luminosities and radii at the measured masses. Future
studies of clusters with more binaries covering a wider range of mass will be important in
addressing this second issue.
We next investigate whether or not there are systematic differences in the different model
sets that might explain the wide range in measured ages. Many variables will affect the
isochrones, among the most significant are adopted α-element enhancement and the relative
abundance distribution over the α-elements, helium abundance, mixing length and convective
overshoot, and the treatment of diffusion and gravitational settling. None of the models
considered include the effects of rotation. Two of these variables ([Fe/H] and α-element
enhancement) are measurable from spectroscopic data, while the others are parameters of
the models.
In the following subsections we explore the sensitivity of the measured ages to [Fe/H],
[α/Fe], diffusion and gravitational settling, and helium abundance.
– 14 –
6.1. Sensitivity to Adopted [Fe/H]
Figure 12 shows the sensitivity of the measured ages to model [Fe/H] for the Dartmouth
and Yonsei–Yale models sets for a range of ±0.05 dex around our adopted value of -0.71 for
each of the three methods of measuring age. These values are interpolated in the models at
a fixed [α/Fe] = +0.4. The plotted values are a weighted average of ages measured for the
primary and secondary. The slopes (in terms of ∆ Gyr per 0.1 dex) are given in Table 10.
In all three cases the slopes for the two model sets are very similar, and Table 10 also gives
the average slopes.
6.2. Sensitivity to α-Element Enhancement
We have estimated the effect of differing α-element enhacement by measuring the ages
of the components of V69 from Dartmouth and Yonsei-Yale isochrones calculated to have
differing α-enhancement in the range 0.2 to 0.4 at a fixed [Fe/H] = -0.71. The results are
shown in Fig. 13, where we plot the weighted average of the ages measured for the primary
and secondary components against model [α/Fe]. As we found for the sensitivity to adopted
model [Fe/H], the slopes (in terms of ∆ Gyr per 0.1 dex) are similar for the two model sets.
The individual values and the averages are listed in Table 10.
6.3. Helium Abundance and Heavy Element Diffusion
At the adopted [Fe/H], the models span a modest but important range in helium abun-
dance (Y ), from Y = 0.243 for the Victoria-Regina models to 0.2555 for the Dartmouth
models. The age estimates from mass-radius, mass-luminosity, and turnoff mass-age rela-
tions for all five model sets are plotted in the upper panel of Fig. 14. An obvious trend in
age with Y can be seen in Fig. 14 as expected. In order to further explore the dependence
of age on Y , stellar evolution tracks were computed by one of us (Dotter) using the code
described by Dotter et al. (2007) for [Fe/H] = -0.71 and [α/Fe] = +0.4 at Y = 0.24, 0.255,
0.27, 0.285, and 0.30 for the measured masses and ±1-σ uncertainties of the primary and
secondary components of V69 (see Table 6). Ages were read from the tracks at the measured
luminosities and radii of the components and plotted as open symbols in Fig. 14 for all but
the Y =0.30 case. The plotted track-based ages are a weighted average of the ages of the
primary and secondary components. These ages are listed in Table 11. The variation in
measured ages with helium abundance for these tracks closely follows that seen for the five
sets of model isochrones.
– 15 –
Some of the scatter seen in the upper panel of Fig. 14 comes from two sources. First,
the Victoria-Regina models are calculated for [α/Fe]=+0.3, while the other four model sets
use +0.4. To account for this we adjusted the Victoria-Regina models by +0.49 Gyr and
+0.67 Gyr for ages determined from the mass-luminosity and mass-radius relations, respec-
tively (see Fig. 13 and Table 10). Second, the Dartmouth and Yonsei-Yale models include the
effects of helium and heavy element diffusion and gravitational settling while the Padova,
Teramo, and Victoria-Regina models do not.5 As a general rule, diffusion will lower the
ages determined from models (Salaris et al. 2000; Chaboyer et al. 2001). To estimate the
importance of diffusion in the measured ages, evolutionary tracks were also calculated by
Dotter for Y = 0.255 but without diffusion. Ages determined from these tracks are plotted
in the upper panel of Fig. 14. It might be expected that the effect of diffusion will depend
on adopted helium abundance, and in the absence of a detailed comparison of diffusion in
multiple model sets we adopt as a rough estimate a correction of -0.7 Gyr as measured from
the Dartmouth tracks with and without diffusion. This correction was applied to the ages
determined from the Padova, Teramo, and Victoria-Regina models. We note that evidence
for diffusion is seen in the globular cluster NGC 6397 (Korn et al. 2007; Lind et al. 2008)
based on a systematic variation of measured metal abundances from the turnoff through
the giant branch. No such evidence was detected in 47 Tuc by Koch & McWilliam (2008),
although the 47 Tuc sample only included one turnoff star.
The corrected ages are plotted in the bottom panel of Fig. 14. There are two conclusions
to be drawn from this figure. First, the measured ages continue to be tightly correlated with
Y , and in particular, remaining residuals from this trend are not strongly related to [α/Fe],
distribution of the α-element enhancements, or the mixing length. This correlation means
that absolute ages of globular clusters based on individual stars in these clusters will be
crucially dependent on an accurate measurement of the helium abundance of the cluster.
This might seem impossible given the accuracy of various methods of determining cluster
helium abundances (see, for example, Sandquist 2000). However, we note that for model
sets with Y . 0.25, the ages measured from mass-luminosity and mass-radius relations differ
systematically, with mass-luminosity relations predicting higher ages. This effect is reversed
for ages measured from Dartmouth tracks for Y = 0.285. If the models are sufficiently
accurate, it should then be possible to constrain both age and Y by requiring that the
mass-luminosity and mass-radius ages be consistent.
5In fact, the situation is more complicated than this because the Dartmouth models employ a treatment
of diffusion that is inhibited in the outer 0.01 M⊙ of the star (Chaboyer et al. 2001). The Teramo models
employ diffusion in the calibration of their solar model but not in models used to generate their metal poor
models and thus their predicted ages are indirectly affected by diffusion.
– 16 –
We investigate the issue of helium abundance further in Fig. 15 where we plot the
measured radii and luminosities of the components of V69 together with Dartmouth tracks
calculated for the masses of the primary and secondary for a range values of the helium
abundance. The measured values of the radii and luminosities of both the primary and
secondary agree to within the uncertainties for Yp = 0.267+0.023−0.029 and Ys = 0.271+0.024
−0.031 with
an average value of Y = 0.269+0.017−0.021. This value can be compared with estimates of the
helium abundance estimated from the population ratio R = NHB/RRGB measured from
ground-based photometry (Y = 0.216+0.013−0.015: Sandquist (2000)) and HST photometry (Y =
0.240 ± 0.015: Salaris et al. (2004)). The agreement is reasonable given the uncertainties in
the models used to calibrate the R method. There are at least two possible ways to improve
upon our estimate. First, Figure 15 shows that our estimate is completely dominated by
errors in the luminosities of the components of V69. Near-infrared eclipse photometry of V69
can be used to improve upon distenace estimates, and hence luminostiy estimates. Second,
discovery of additional similar systems to V69 in 47 Tuc will help improve improve upon
distance, age, and helium abundance measurements for this cluster.
Casagrande et al. (2007) derived bolometric luminosities and effective temperatures
from multi-band photometry and the infrared flux method for a sample of 86 K-dwarfs.
They compared these values to Padova stellar isochrones to estimate the He to metal en-
richment ratio (∆Y/∆Z) in the solar neighborhood. Although their primary interest was
to measure ∆Y/∆Z, Casagrande et al. found that, in an absolute sense, the implied He
content of the most metal-poor stars was as low as Y∼0.1–in marked disagreement with
canonical values for the primordial He content (Salaris et al. 2004; Spergel et al. 2007). In
contrast, we find no indication that current stellar evolution models are unable to reproduce
the mass-radius relationship of V69.
6.4. The Absolute Age of 47 Tuc
As mentioned above, a measurement of the absolute age of 47 Tuc will have random
and systematic errors. Our best age estimates derived from the model isochrones have
statistical errors of about 0.25 Gyr for mass-radius relations (see Table 9). Outside of
remaining systematic errors arising from different model parameters, the main systematic
errors arise from the empirical estimates for metallicity and α-element enhancement. We
estimate the errors in both of these measurements to be 0.10 dex, and following the results
in Table 10 we adopt a systematic error of 0.85 Gyr, the quadratic sum of the contributions
from metallicity and α-element enhancement. This leads to an absolute age estimate of
11.25±0.21(random)±0.85(systematic) Gyr from Dartmouth isochrones adopting Y = 0.255,
– 17 –
[Fe/H] = -0.71, and [α/H] = 0.4. Note that the age estimate from the Teramo isochrones
is very similar when a correction for diffusion has been applied (see Table 9 and Fig 14),
as are the ages derived from mass-luminosity relations for both the Dartmouth and Teramo
models. If we adopt a helium abundance of Y = 0.27 then the Dartmouth tracks indicate
an age of 10.21±0.21±0.85 Gyr.
This age estimate is consistent with other recent estimates of the age of 47 Tuc. For
example, Gratton et al. (2003) find an age of 11.2 ± 1.1 by fitting model isochrones with no
diffusion to the cluster color-magnitude diagram, and an age of 10.8 ± 1.1 Gyr for models
including diffusion. Salaris & Weiss (2002) derive an age of 10.7 ± 1.0, also from isochrone
fitting. Zoccali et al. (2001) obtain an age of 13± 2.5 Gyr by fitting the white-dwarf cooling
sequence. Grundahl et al. (2002) concluded that the age of 47 Tuc is ”slightly below 12 Gyr”
based on analysis of Stromgren photometry and isochrone fitting. Our determination of the
age of 47 Tuc based on observations of V69 has a small statistical error compared to that
derived from isochrone fitting.
7. DISCUSSION AND SUMMARY
To the best of our knowledge, these measurements of the masses and radii of the com-
ponents of V69 are the first such high accuracy (better than 1%) measurements to be made
for Population II stars. The binary is a member of the globular cluster 47 Tuc and so the
determination of its distance and age applies to the cluster as well. We obtained a distance
modulus of (m − M)V = 13.35 ± 0.08. The main source of error in the distance estimate
is the calibration dependent estimate of effective temperatures which we used to derive the
bolometric luminosities for the component stars.
A comparison of the measured masses, luminosities, and radii of the components to
stellar evolution models suggests that the age of the system and hence the globular cluster
47 Tuc can be measured to a statistical accuracy of about 0.25 Gyr. However, it is important
to understand the assumptions that go into any one model. In particular, the derived
ages are very sensitive to the adopted helium abundance. We derive an age for 47 Tuc of
11.25±0.21±0.45 Gyr for a helium abundance Y = 0.255 using Dartmouth model isochrones.
All models give similar ages when the effects of helium abundance are taken into account.
Comparison of Dartmouth evolutionary tracks calculated for the measured masses of
the primary and secondary indicate that the helium abundance can be measured to an
accuracy of about 0.03 for each of the components. We estimate a helium abundance of
Y = 0.269+0.017−0.021 for 47 Tuc. The measured masses, radii, and luminosities of the components
– 18 –
of V69 are consistent with Dartmouth models assuming [Fe/H] = -0.71, [α/Fe] = + 0.4, and
Y = 0.27.
The radii of both stars are known with high accuracy, and it is therefore possible to
obtain a more accurate and robust distance determination based on the surface brightness
method (Barnes & Evans 1976; Lacy 1977; Thompson et al. 2001). The empirical calibration
of surface brightness relations for dwarf and subgiant stars is improving (Di Benedetto 1998;
Kervella et al. 2004; Buermann 2006), and it is reasonable to imagine that a distance accurate
to a few per cent can be measured with accurate radii and (V −K) colors. We are in the
process of collecting near IR eclipse profile photometry of both V69 and V228 (Kaluzny et al.
2007b) in order to measure the distances to these two binary stars in this way. These data
will improve the estimates of the bolometric luminosities of the components and lead to a
more accurate measurement of the helium abundance and hence the absolute age of 47 Tuc.
Finally, we note that contributions to the errors in the radii are dominated by the
photometric solution. Given the large inclination, the errors in the masses are completely
dominated by the orbital solution. An identical doubling of the existing set of radial ve-
locity observations leads to a 33% improvement in the mass estimates, and a subsequent
improvement in the age estimates. The system is relatively bright, and the prospects are
good that a substantial improvement in the measured masses can be achieved with further
radial velocity observations.
JK, WP and WK were supported by grant N203 379936 from the Ministry of Science and
Higher Education, Poland. Research of JK is also supported by the Foundation for Polish
Science through the grant MISTRZ. IBT is supported by NSF grant AST-0507325. Support
from the Natural Sciences and Engineering Council of Canada to SMR is acknowledged with
gratitude. It is a pleasure to thank John Southworth and Willy Torres for sharing software
with us. Jay Anderson kindly measured the proper motion of V69. IBT acknowledges useful
conversations with Chris Burns, Andy McWilliam, and George Preston. This research used
the facilities of the Canadian Astronomy Data Centre operated by the National Research
Council of Canada with the support of the Canadian Space Agency. We dedicate this series
of papers to the memory of our colleague Bohdan Paczynski.
Appendix
In their catalog of 47 Tuc variables Weldrake et al. (2004) included a potentially inter-
esting detached binary named V39. The object has an orbital period of 4.6 d and is located
about 0.30 mag to the red of the cluster main-sequence on the V/V − I plane. We have
– 19 –
obtained BV images of the V39 field using the du Pont telescope. These images show that
V39 is a close visual pair of stars separated by 1.5 arcsec. The brighter component has
V ≈ 18.1 and B−V ≈ 0.99 and is located on a V/B−V CMD among the Small Magellanic
Cloud (SMC) asymptotic giant branch stars. The fainter component has V ≈ 19.2 and
B − V ≈ 0.08 and is a candidate SMC upper main sequence star. Given the orbital period
of V39 we propose that it is the fainter component of the blend which is the eclipsing binary.
REFERENCES
Alves-Brito, A., et al. 2005, A&A, 435, 657
Andersen, J. 1991, A&AR, 3, 91
Anderson, J. & King, I. R. 2003, AJ, 126, 772
Bahcall, J. N., Basu, S., Pinsonneault, M., & Serenelli, A. M. 2005, ApJ, 618, 1049
Barnes, T. G. & Evans, D. S. 1976, MNRAS, 174, 489
Bernstein, R., Shectman, S. A., Gunnels, S. M., Mochnacki, S., & Athey, A. E. 2003, Instru-
ment Design and Performance for Optical/Infrared Ground-based Telescopes. Edited
by Iye, Masanori; Moorwood, Alan F. M. Proceedings of the SPIE, 4841, 1694
Blake, C. H., Torres, G., Bloom, J. S., & Gaudo, B. S. 2008, ApJ, 684, 635
Bono, G. et al. 2008, ApJ, 686, L87
Boyajian, T. B. et al. 2008, ApJ, 683, 2008
Brown, J. A. & Wallerstein, G. W. 1992, AJ, 104, 1818
Buermann, K. 2006, A&A, 460, 783
Carretta, E. & Gratton, R. 1997, A&AS, 121, 95
Carretta, E., Bragaglia, A., Gratton, R., & Lucatello, S. 2009, preprint (arXiv:0909.2941)
Casagrande, L., Flynn, C., Portinari, L., Girardi, L., & Jimenez, R. 2007, MNRAS, 382,
1516
Chaboyer, B., Fenton, W. H., Nelan, J. E., Patnaude, D. J., & Simon, F. E. 2001, ApJ, 562,
521
– 20 –
Clausen, J. V., Torres, G., Bruntt, H., Andersen, J. Nordstrom, Stefanik, R. P., Latham, D.
W., & Southworth, J. 2008, A&A, 487, 1095
Coelho, P., Barbuy, B., Melendez, J., Schiavon, R. P., &Castilho, B. V. 2005, A&A, 443,
735
Di Benedetto, G. P. 1998, A&A, 339, 858
Dotter, A., Chaboyer, B., Jevremovic, D., Baron, E., Ferguson, J.W., Sarajedini, A., &
Anderson, J. 2007, AJ, 134, 376
Etzel, P. B. 1981, in Photometric and Spectroscopic Binary Systems, ed. E. B. Carling &
Z. Kopal, 111–120
Gebhardt, K., Pryor, C., Williams, T. B., & Hesser, J. E. 1995, AJ, 110, 1699
Gratton, R.G., Bragaglia, A., Carretta, E., Clementini, G., Desidera, S., Grundahl, F., &
Lucatello, S. 2003, A&A, 408, 529
Grevesse, N. & Noel, A. 1993, in Origin and Evolution of the Elements, ed. N. Prantos, E.
Vangioni-Flam, & M. Casse(Cambridge: Cambridge Univ. Press), 15
Grevesse, N. & Sauval, A. J. 1998, Space Sci. Rev., 85, 161
Grundahl, F., Stetson, P. B., & Anderson, M. I. 2002, A&A, 395, 481
Grundahl, F., Clausen, J. V., Hardis, S., &Frandsen, 2008, A&A, 492, 171
Harris, W. E. 1996, AJ, 112, 1487
Kaluzny, J., Kubiak, M., Szymanski, M., Udalski, A., Krzeminski, W., Mateo M., & Stanek,
K. Z., 1998, A&AS, 128, 19
Kaluzny, J., Thompson, I.B., Krzeminski, W., Olech, A., Pych, W., & Mochejska, B. 2002,
in ASP Conf. Ser. 265, Omega Centauri, A Unique Window into Astrophysics, ed. F.
can Leewen, J. D. Hughes, & G. Piotto (San Francisco: ASP)
Kaluzny, J., et al. 2005, in AIP Conf. Proc. 752, Stellar Astrophysics with the World’s
Largest Telescopes, Ed. J. Miko lajewska & A. Olech (Melville, New York), 70
Kaluzny, J., Pych, W., Rucinski, S., & Thompson, I. B. 2006, Acta Astron., 56, 237
Kaluzny, J., Rucinski, S. M., Thompson, I. B., Pych, W., Krzeminski, W., 2007a, AJ, 133,
2457
– 21 –
Kaluzny, J., Thompson, I. B., Rucinski, S. M., Pych, W., Stachowski, G., Krzeminski, W.,
Burley, G. S. 2007b, AJ, 134, 541
Kaluzny, J., Thompson, I. B., Rucinski, S. M., Krzeminski, W., 2008, AJ, 136, 400
Kelson D. D. 2003, PASP, 115, 688
Kervella, P., Thevenin, F., Di Folco, E. Segransan, D. 2004, A&A, 426, 297
Koch, A. & McWilliam, A. 2008, AJ, 135, 1551
Korn, A. J., Grundahl, F., Richard, O., Mashonkina, L., Barklem, P.S., Collett, R., Gustafs-
son, B., & Piskunov, N. 2007, ApJ, 671, 402
Kim, Y.-C., Demarque, P., Yi, S. K., & Alexander, D. R. 2002, ApJS, 143,499
Kraft, R. P. & Ivans, I. I. 2003, PASP, 115, 143
Landolt, A. U. 1992, AJ, 104, 340
Lacy, C. H. S. 1977, ApJ, 213, 458
Lacy, C. H. S., Torres, G., Claret, A., & Vaz, L. P. R. 2005, AJ, 130, 2838
Lacy, C. H. S., Torres, G., & Claret, A. 2008, AJ, 135, 1757
Lind, K., Korn, A. J., Barklem, P.S., & Grundahl, F. 2008, A&A, 490, 777
Lopez-Morales, M., Orosz, J. A., Shaw, J. S., Havelka, L., Arevalo, M. J., McIntyre, T., &
Lazaro 2006, preprint (astro-ph/0610225)
Lopez-Morales, M. 2007, ApJ, 660, 732
McLaughlin, D. E., Anderson, J., Meylan, G., Gebhardt, K., Pryor, C., Minniti, D., &
Phinney, S. 2006, ApJS, 166, 249
Meibom, S. 2009, preprint (arXiv:0903.3566)
Neckel, H. 1986, A&A, 159, 175
Norris, J. E. & Da Costa, G. 1995, ApJ, 447, 680
Paczynski, B. 1997, in Space Telescope Science Institute Series, The Extragalactic Distance
Scale, ed. M. Livio (Cambridge University Press), p. 273
Percival, S. M., Salaris, M., van Vyk, F., & Kilkenny, D. 2002, ApJ, 573, 174
– 22 –
Pietrinferni, A., Cassisi, S., Salaris, M., & Castelli, F. 2006, ApJ, 642, 797
Popper, D. M. 1980, ARA&A, 18, 115
Popper, D. M. & Etzel, P. B. 1981, AJ, 86, 102
Prsa, A. & Zwitter, T. 2005, ApJ, 628, 426
Ramirez, I., & Melendez, J. 2005, ApJ, 626, 465
Salaris, M., Groenewegen, M. A. T., & Weiss, A. 2000, A&A, 355, 299
Salaris, M. & Weiss, A. 2002, A&A, 388, 492
Salaris, M., Riello, M., Cassisi, S., & Piotto, G. 2004, A&A, 420, 911
Salaris, M., Held, E.V., Ortolani, S., Gullieszik, M., & Momany, Y. 2007, A&A, 476, 243
Salasnich, B., Girardi, L., Weiss, A., & Chiosi, C. 2000, A&A, 361, 1023
Sandquist, E. L. 2000, MNRAS,313, 571
Sarajdeni, A., et al. 2007, AJ, 133, 1658
Schlegel, D., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525
Serenelli, A. M., Basu, S., Ferguson, J. W., &Asplund, M. 2009, preprint (arXiv:0909.2668)
Southworth, J., Maxted, P. F. L., & Smalley, B. 2004a, MNRAS, 351, 1277
Southworth, J., Zucker, S., Maxted, P. F. L., & Smalley, B. 2004b, MNRAS, 355, 986
Southworth, J., Bruntt, H., & Buzasi, D. L. 2007, A&A, 467, 1215
Spergel, D. N. et al. 2007, ApJS, 170, 377
Stetson, P. B. 1987, PASP, 99, 191
Stetson, P.B. 1990, PASP, 102, 932
Stetson, P. B. 2000, PASP, 112, 925
Thompson, I. B., Kaluzny, J., Pych, W., Burley, G. S., Krzeminski, W., Paczynski, B.,
Persson, S. E., & Preston, G. W. 2001, AJ, 121, 3089
Torres, G., Boden, A. F., Latham, D. W., Pan, M., & Stefanik, R. P., AJ, 124, 1716
– 23 –
VandenBerg, D.A., & Clem, J.L. 2003, AJ, 126, 778
VandenBerg, D. A., Bergbusch, P. A., & Dowler, P. D. 2006, ApJS, 162, 375
van Hamme, W. 1993, AJ, 106, 2096
Weldrake, D. T. F., Sackett, P. D., Bridges, T. J., & Freeman, K. C. 2004, AJ, 128, 736
Wilson, R. E., & Devinney, E. J. 1971, ApJ, 166, 605
Wilson, R. E. 1979, ApJ, 234, 1054
Worthey, G., & Lee, H.-C. 2006, preprint (astro-ph/0604590)
Zoccali, M. et al. 2001, ApJ, 553, 733
Zucker, S. & Mazeh, T. 1994, ApJ, 420, 806
This preprint was prepared with the AAS LATEX macros v5.2.
– 24 –
Table 1. BV Photometry of V69 at Minima and Quadrature
Phase V B B − V
Max 16.836(1) 17.384(2) 0.548(2)
Min I 17.507(3) 18.083(12) 0.576(12)
Min II 17.461(3) 18.011(14) 0.550(14)
– 25 –
Table 2. Radial Velocity Observations of V69
HJD Vp Vs Phase
(- 2400000) km s−1 km s−1
53183.88089 −57.33 24.10 0.173
53201.93116 21.23 −55.92 0.784
53206.81454 −5.58 −28.40 0.950
53206.93425 −7.00 −27.15 0.954
53210.82680 −41.09 8.46 0.085
53210.93896 −42.27 9.98 0.089
53271.78367 −53.96 20.75 0.149
53274.77303 −59.04 26.63 0.250
53280.65021 −23.44 −9.68 0.449
53281.65012 −14.50 −18.34 0.483
53282.69872 −7.02 −26.32 0.519
53521.93280 12.54 −45.81 0.617
53580.91364 11.43 −46.41 0.614
53581.85305 16.04 −50.05 0.646
53582.85552 19.69 −53.69 0.680
53584.80006 22.23 −56.51 0.745
53585.79822 21.34 −55.77 0.779
53631.71181 −50.28 16.17 0.334
53633.74856 −36.21 1.48 0.402
53634.76492 −27.12 −6.54 0.437
53889.93336 −37.97 6.09 0.075
53891.93240 −52.13 20.77 0.143
53892.93208 −56.74 24.68 0.177
53893.93261 −59.42 26.95 0.210
53935.93349 14.75 −48.09 0.632
53937.92995 21.18 −55.15 0.700
53938.92894 22.39 −56.40 0.734
– 26 –
Table 3. Orbital Parameters for V69
Parameter Value
P (days) 29.53975a
T0 (HJD-240 0000) 2453237.8421a
γ (km s−1) -16.71± 0.05
Kp (km s−1) 41.03±0.13
Ks (km s−1) 41.86±0.09
e 0.0563±0.0011
ω (deg) 149.15±1.86
σp (km s−1) 0.50
σs (km s−1) 0.33
Derived quantities:
A sin i (R⊙) 48.301±0.095
Mp sin3 i (M⊙) 0.8762±0.0048
Ms sin3 i (M⊙) 0.8588±0.0060
aEphemeris adopted from photome-
try.
– 27 –
Table 4. Results of the Light Curve Analysis for V69 Obtained with the PHOEBE Code
Parameter
i (deg) 89.771± 0.009
Tp (K) 5945
Ts (K) 5959±3
e 0.0556±0.0002
ω (deg) 149.72±0.25
rp 0.02727±0.00005
rs 0.02408±0.00008
(Ls/Lp)V 0.7870±0.0012
(Ls/Lp)B 0.7889±0.0015
σrms (V ) (mmag) 10.0
σrms (B) (mmag) 14.5
– 28 –
Table 5. Results of the Light Curve Analysis for V69 Obtained with the JKTEBOP Code
Parameter V B
Adjusted quantities
rp + rs 0.05166± 0.00009 0.05151±0.00021
k = rp/rs 0.8836 ± 0.0072 0.8961±0.0199
i (deg) 89.768± 0.012 89.769±0.028
J 1.0146±0.0031 1.0022±0.0091
e 0.0567±0.0008 0.0586±0.0025
ω (deg) 147.7±1.3 145.2±3.4
Other quantities
rp 0.02722±0.00009 0.02717±0.00026
rs 0.02405±0.00012 0.02435±0.00033
Ls/Lp 0.792±0.011 0.802± 0.028
σrms (mmag) 10.0 14.5
– 29 –
Table 6. Absolute Parameters for V69
Parameter Value
A (R⊙) 48.30±0.12
Mp (M⊙) 0.8762±0.0048
Ms (M⊙) 0.8588±0.0060
Rp (R⊙) 1.3148±0.0051
Rs (R⊙) 1.1616±0.0062
Tp (K) 5945±150
Ts (K) 5959±150
Lbolp (L⊙) 1.94 ±0.21
Lbols (L⊙) 1.53 ±0.17
MVp (mag) 4.12 ±0.11
MVs (mag) 4.38 ±0.12
log gp(cm s−2) 4.143 ±0.003
log gs(cm s−2) 4.242 ±0.003
– 30 –
Table 7. Comparison of Distance Determinations to 47 Tuc
Author Method (m−M)V
Gratton et al. (2003) Main Seq. fitting 13.50±0.08
Percival et al. (2002) Main Seq. fitting 13.37±0.11
Grundahl et al. (2002) Main Seq. fitting 13.33 ± 0.04 ± 0.1
Zoccali et al. (2001) white dwarfs 13.27±0.14
McLaughlin et al. (2006) kinematics 13.15 ± 0.13
Salaris et al. (2007) IR luminosity of HB 13.18 ± 0.03 ± 0.04
Kaluzny et al. (2007b) Eclipsing binary 13.40 ± 0.07
Bono et al. (2008) Tip RG branch 13.32 ± 0.09
Bono et al. (2008) IR lum. of RR Lyrae stars 13.47 ± 0.11
This paper Eclipsing binary 13.35 ± 0.08
– 31 –
Table 8. Parameters of Stellar Evolution Models
Parameter Dartmouth Padova Teramo Victoria Yonsei-Yale
[Fe/H] −0.71 -0.70 −0.70 −0.705 −0.71
[α/Fe] +0.40 0.40 +0.40 +0.30 +0.40
X 0.738010 0.742000 0.736000 0.750250 0.748031
Y 0.255000 0.250000 0.256000 0.243000 0.244646
Z 0.006490 0.008000 0.008000 0.006750 0.007323
[O/Fe] 0.40 0.50 0.50 0.30 0.40
mixing length 1.938 1.680 1.913 1.890 1.743
(Z/X)solar 0.023 0.024 0.024 0.024 0.024
–32
–
Table 9. Ages in Gyr for the Components of V69 Derived from Model Isochrones
Method Component Dartmouth Padova Teramo Victoria-Regina Yonsei-Yale
Mass - Luminosity primary 11.11+0.54/-0.58 12.39+0.55/-0.62 11.93+0.55/-0.62 13.62+0.57/-0.66 13.02+0.54/-0.62
secondary 11.12+0.80/-0.93 12.46+0.80/-0.89 12.04+0.76/-0.93 13.90+0.68/-0.89 13.16+0.72/-0.81
average 11.11+0.45/-0.49 12.41+0.45/-0.51 11.97+0.45/-0.52 13.74+0.44/-0.53 13.07+0.43/-0.49
Mass - Radius primary 11.25±0.26 12.26±0.29 11.97±0.31 13.32±0.33 12.63±0.29
secondary 11.25±0.37 12.34±0.40 12.16±0.42 13.54±0.46 12.70±0.39
average 11.25±0.21 12.29±0.24 12.04±0.25 13.39±0.27 12.65±0.23
Turnoff Mass primary 10.91±0.24 11.81±0.27 11.19±0.25 12.50±0.29 11.57±0.26
secondary 11.75±0.31 12.79±0.37 12.15±0.35 13.58±0.40 12.52±0.34
– 33 –
Table 10. Effect of [Fe/H] and [α/Fe] on Measured Age
Method Model [Fe/H] [α/Fe]
∆Gyr / 0.1 dex ∆Gyr / 0.1 dex
Mass - Luminosity Dartmouth 1.09 0.67
Yonsei-Yale 1.03 0.67
average 1.06 0.67
Mass - Radius Dartmouth 0.73 0.44
Yonsei-Yale 0.79 0.54
average 0.76 0.49
Turnoff Mass Dartmouth 0.68 0.43
Yonsei-Yale 0.78 0.37
average 0.73 0.40
–34
–
Table 11. Ages in Gyr for the Components of V69 Derived from Dartmouth Tracks
Method Component Y = 0.24 Y = 0.255 Y = 0.255a Y = 0.27 Y = 0.285
Luminosity primary 13.51+0.64/-0.66 11.77+0.56/-0.65 12.27+0.56/-0.65 10.20+0.52/-0.63 8.77+0.51/-0.62
secondary 13.71+0.76/-0.87 11.85+0.72/-0.86 12.40+0.76/-0.93 10.16+0.71/-0.90 8.60+0.74/-0.93
average 13.59+0.49/-0.53 11.80+0.44/-0.52 12.32+0.45/-0.53 10.19+0.42/-0.52 8.72+0.42/-0.52
Radius primary 12.97±0.32 11.53±0.28 12.32±0.30 10.23±0.25 9.10±0.22
secondary 12.98±0.45 11.48±0.40 12.49±0.43 10.13±0.36 8.94±0.32
average 12.97±0.26 11.51±0.23 12.38±0.25 10.21±0.21 9.05±0.18
aTracks calculated without the effects of diffusion.
– 35 –
Fig. 1.— Plot of the color and magnitude residuals for the standard stars observed on the
night of 2007 August 19.
– 36 –
Fig. 2.— The phased BV light curves of V69.
– 37 –
Fig. 3.— Radial velocity observations of V69. Filled symbols represent data for the primary
and open symbols are for the secondary.
– 38 –
B-V
V
Fig. 4.— Position of V69 in the BV CMD for 47 Tuc. The square gives the position for the
combined light and the triangles show the positions of each of the individual components.
– 39 –
Fig. 5.— The residuals of the fits to the light curve obtained with JKTEBOP for the
primary (bottom) and secondary eclipse (top). The residuals for B are offset by 0.15 mag
for clarity.
– 40 –
Fig. 6.— The masses, luminosities, and radii of the components of V69 are compared to
isochrones based on the Dartmouth models. For each panel the isochrones are plotted in
steps of 0.5 Gyr, with the lowest and highest age isochrones labeled.
– 41 –
Fig. 7.— The masses, luminosities, and radii of the components of V69 are compared to
isochrones based on the Padova models. For each panel the isochrones are plotted in steps
of 0.5 Gyr, with the lowest and highest age isochrones labeled.
– 42 –
Fig. 8.— The masses, luminosities and radii of the components of V69 are compared to
isochrones based on the Teramo models. For each panel the isochrones are plotted in steps
of 0.5 Gyr, with the lowest and highest age isochrones labeled.
– 43 –
Fig. 9.— The masses, luminosities, and radii of the components of V69 are compared to
isochrones based on the Victoria-Regina models. For each panel the isochrones are plotted
in steps of 0.5 Gyr, with the lowest and highest age isochrones labeled.
– 44 –
Fig. 10.— The masses, luminosities, and radii of the components of V69 are compared to
isochrones based on the Yonsei-Yale models. For each panel the isochrones are plotted in
steps of 0.5 Gyr, with the lowest and highest age isochrones labeled.
– 45 –
Fig. 11.— Age as a function of turnoff mass for the Dartmouth (solid line), Teramo (dotted
line), Yonsei-Yale (long dashed line), Padova (dot-dashed line), and Victoria-Regina (short
dashed line) models. The masses for the primary and secondary components of V69 are
plotted as solid vertical lines, with one-sigma errors represented by the shaded areas about
the component masses.
– 46 –
Fig. 12.— Age as a function of [Fe/H] for Dartmouth (filled symbols) and Yonsei-Yale (open
symbols) models. Ages are determined from mass-luminosity relations (bottom panel), mass-
radius relations (middle panel), and turnoff mass - age relations (upper panel).
– 47 –
Fig. 13.— Age as a function of α-element enhancement for Dartmouth (filled symbols) and
Yonsei-Yale (open symbols) models. Ages are determined from mass-luminosity relations
(bottom panel), mass-radius relations (middle panel), and turnoff mass - age relations (upper
panel).
– 48 –
Fig. 14.— Age as a function of helium abundance as measured from Victoria-Regina, Yonsei-
Yale, Padova, Dartmouth, and Teramo isochrones (solid symbols plotted left to right, upper
panel). Ages are determined from mass-luminosity relations (circles), mass-radius relations
(squares), and turnoff mass - age relations (triangles). Open symbols represent ages measured
from Dartmouth tracks, see text for details. Ages from two sets of Dartmouth tracks are
plotted for Y = 0.255, models including helium and heavy element diffusion (lower ages)
and models without diffusion (higher ages). The plotted points have been offset in helium
abundance for clarity, the circles represent the correct helium abundance for any one model
set. Bottom panel: Ages corrected for the effects of diffusion (Victoria-Regina, Padova, and
Teramo models). The Victoria-Regina ages have been additionally corrected to bring the
α-element enhancement to +0.4.
– 49 –
Fig. 15.— Dartmouth tracks calculated for the measured masses of the primary (left panel)
and secondary (right panel). In each case the tracks are calculated for Y = 0.24 through
Y = 0.30. The measured one sigma limits on mass are plotted as dotted lines for the Y = 0.27
tracks. The measured values of radius and luminosity for the components of V69 are both
consistent with Y = 0.27 with a one sigma range of 0.03.